INDEX

Explanations

mentions of the word "mother."

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of the word "mother."

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ickr

-0.84

 Flavoring

-0.81

EY

-0.81

raviolet

-0.78

NRS

-0.74

okers

-0.71

OFF

-0.69

vernment

-0.69

Vert

-0.67

OPLE

-0.67

POSITIVE LOGITS

hood

1.10

hesis

1.04

mother

1.00

heses

0.88

ship

0.88

hetically

0.86

wife

0.85

 mother

0.83

ma

0.82

stones

0.80

Activations Density 0.016%

mentions of the word "mother."

repeated mentions of the word "mother."

No Comments

No Known Activations